Link Prediction on Social Networks Based on Centrality Measures

Abstract

In order to understand and compare different social networks, the role of nodes, edges, and their relative importance needs to be investigated. The importance measures of nodes and edges within networks are useful for many social network analysis tasks and are known as centrality measures. These centrality measures allow analysts to investigate network structure, identify influential users, predict future connections, and perform many other related tasks. This chapter presents a study of centrality measures to predict future links in the network. We also choose four different metrics of Recall, Precision, AUPR, and AUC and evaluate the performance of different centrality measures corresponding to the link prediction problem on real-world social networks. © 2022, The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd.